The safety of treatment with digoxin in patients with acute myocardial infarction (MI) was investigated in 584 hospital survivors of MI. All patients were examined by radionuclide ventriculography, with determination of left ventricular ejection fraction (LVEF), close to the time of discharge. Clinical data were collected on admission. All patients were followed up with regard to death (median 6.2 years, range 3.9-7.8 years). Patients treated with digoxin (N = 172 (29%) were older (median 66 vs 59 years; (P < 0.001), had a higher incidence of diabetes (13% vs 7%; P = 0.025), and a lower LVEF (0.33 vs 0.49; P < 0.001). As expected, clinical heart failure was more frequent among them (84% vs 14%; P < 0.001), than in patients not receiving digoxin. The 1- and 5-year mortality of patients treated with digoxin was 38% and 74% compared to 8% and 26% in patients not receiving digoxin (P < 0.001). The increased risk associated with digoxin therapy remained statistically significant when patients were stratified according to the presence or absence of heart failure or atrial fibrillation/flutter during hospitalization, or to LVEF above or below 0.45 at discharge. In a proportional hazard model including age, LVEF, diabetes mellitus, heart failure, atrial fibrillation or flutter, ventricular fibrillation, gender, dose of furosemide at discharge and calcium antagonists and digoxin treatment as covariates, digoxin was independently associated with an increased risk of death (relative risk 1.8 (95% confidence limit 1.2-2.5)). We conclude that administration of digoxin may be harmful in hospital survivors of MI.